Survey operations - 19 May 2026

Altitude, coverage, and revisit have numbers

EmberScope carries a first-order survey model for comparing centimetre-scale sampling against area coverage, loaded endurance, fleet size, dwell time, and alert latency.

Compare scenarios Read mission fit Read detection pipeline Read calibration strategy Read field-test plan

Current model

The 120 m provisional value is coverage-biased; 80 m is the current balanced case.

5.9 cm Current 120 m provisional value GSD with the 18 degree horizontal field of view.

38.0 m Ground swath at 120 m before the 30 percent overlap route-spacing allowance.

6.5 min Nominal 100 ha revisit for a synchronized 10-drone cycle at the 120 m provisional value.

7.0 s Modelled first alert-packet latency after frame capture, separate from revisit time.

What changed

Survey geometry is a camera-level design input, not a loose requirement.

This dated model ties the preferred 640 x 512, 12 um detector branch and 18 degree horizontal field of view to altitude, speed, overlap, loaded endurance, fleet size, dwell, and revisit interval.

The model deliberately stops before detection probability. It gives radiometry and end-to-end simulation work explicit operating rows instead of treating the drone survey path as background context.

Scenario results

Close inspection, balanced sweep, and fleet coverage behave very differently.

Scenario	GSD	Swath	Spacing	Dwell	Area rate	100 ha revisit
35 m inspection/reference mapping	1.7 cm	11.1 m	1.1 m	4.4 s	0.6 ha/h per drone	9395.5 min
80 m candidate sweep, one drone	4.0 cm	25.3 m	17.7 m	2.5 s	40.9 ha/h per drone	146.8 min
120 m current provisional value, one drone	5.9 cm	38.0 m	26.6 m	2.5 s	92.0 ha/h per drone	65.2 min
120 m current provisional value, 10-drone cycle	5.9 cm	38.0 m	26.6 m	2.5 s	92.0 ha/h per drone	6.5 min

Engineering readout

The model separates overflight alert speed from rediscovery speed.

120 m is not settled

It gives useful area rate, but 5.9 cm GSD is close to the stated few-centimetre target scale before blur, diffraction, calibration error, atmosphere, and noise are included.

35 m is a validation mode

The high-overlap case gives strong sampling and dwell, but route spacing collapses to about 1.1 m, so it cannot stand in for broad-area search.

80 m is the next useful row

The balanced case keeps GSD near 4.0 cm while still covering about 40.9 ha/h per drone after the current operations allowance.

Fleet size changes revisit

Ten drones do not change a single pass, but they can reduce a 100 ha provisional value revisit from about 65 minutes to about 6.5 minutes if operations can support parallel lanes.

Next decisions

Stakeholder inputs map directly to model parameters.

The next closure items are search-block size, acceptable revisit time, minimum GSD after optical and radiometric losses, whether a 10-drone cycle is operationally real, and the false-alarm budget per flight minute and per hectare.

Those values will decide whether the radiometry work should treat 80 m or 120 m as the primary case, and whether the final optical field of view should prioritize narrow high-resolution detection or wider faster sweep.